Extraction of Unexpected Precedence between Items in Consumer Price Index

Bibliographic Information

Other Title
  • 物価指数を題材とした品目間の意外な先行性の抽出手法

Search this article

Description

When analyzing marketing data, if we can easily find precedence between A and B that are semantically far from each other, people cannot think of such relationship so that it helps us consider a sales plan focusing on A and B before anyone else do that. Therefore, this paper proposes to define semantic similarity between two items, quantify precedence between two time series data corresponding to these items by using cross correlation, and clarify the gap between the similarity and the precedence as unexpectedness. The proposed method is evaluated with consumer price index data in terms of regression and predictivity.

Journal

Details 詳細情報について

Report a problem

Back to top